[lug] Etch AMD kernels not working

[Gary Hodges]

Gary Hodges wrote:
> David L. Anselmi wrote:
>> Gary Hodges wrote:
>> [...]
>>> A lot of self replies, but I have some new information.  It seems the 
>>> issue is SMP, not a particular kernel flavor.
>>
>> No surprise.
>>
>> Does this help:
>>
>> http://lkml.org/lkml/2006/9/26/303
> 
> Nice link.  Thank you.  In an effort to keep things simple, I was 
> compiling bare bones kernels.  I never included ACPI.  The link you 
> provided indicates it is necessary.  As this is a reasonably busy 
> machine I have to schedule maintenance.  I'll report back results when I 
> get them.

Last week I was able to try a kernel compiled with ACPI.  The machine 
successfully booted three or four consecutive times without trouble, so 
we decided to let it run and see what happened.  After about four hours 
of use we declared it "fixed."  Then about five minutes later it locked 
up again.

I've now updated about a dozen machines from Sarge to Etch.  All the 
Intel (all single CPU) machines updated smoothly.  The only two AMD 
machines I updated both had issues.  One of those had a likely hardware 
problem so I'll give it a pass.  The other (the subject of my many 
posts) has never been stable running as SMP, though it ran for years as 
SMP with Sarge installed.

Not sure what we'll do next.  Maybe update the hardware, or try a 
different OS, but I'm about finished trying to get Etch to run on it as 
it stands now.